ceph-csi/internal/rbd/rbd_attach.go

617 lines
18 KiB
Go
Raw Normal View History

/*
Copyright 2018 The Ceph-CSI Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package rbd
import (
"context"
"encoding/json"
"fmt"
"os"
"slices"
"strings"
"time"
"github.com/ceph/ceph-csi/internal/util"
"github.com/ceph/ceph-csi/internal/util/log"
"github.com/container-storage-interface/spec/lib/go/csi"
"k8s.io/apimachinery/pkg/util/wait"
)
const (
rbdTonbd = "rbd-nbd"
moduleNbd = "nbd"
accessTypeKRbd = "krbd"
accessTypeNbd = "nbd"
rbd = "rbd"
// Output strings returned during invocation of "rbd unmap --device-type... <imageSpec>" when
// image is not found to be mapped. Used to ignore errors when attempting to unmap such images.
// The %s format specifier should contain the <imageSpec> string
// NOTE: When using devicePath instead of imageSpec, the error strings are different.
rbdUnmapCmdkRbdMissingMap = "rbd: %s: not a mapped image or snapshot"
rbdUnmapCmdNbdMissingMap = "rbd-nbd: %s is not mapped"
rbdMapConnectionTimeout = "Connection timed out"
defaultNbdReAttachTimeout = 300 /* in seconds */
defaultNbdIOTimeout = 0 /* do not abort the requests */
// The default way of creating nbd devices via rbd-nbd is through the
// legacy ioctl interface, to take advantage of netlink features we
// should specify `try-netlink` flag explicitly.
useNbdNetlink = "try-netlink"
// `reattach-timeout` of rbd-nbd is to tweak NBD_ATTR_DEAD_CONN_TIMEOUT.
// It specifies how long the device should be held waiting for the
// userspace process to come back to life.
setNbdReattach = "reattach-timeout"
// `io-timeout` of rbd-nbd is to tweak NBD_ATTR_TIMEOUT. It specifies
// how long the IO should wait to get handled before bailing out.
setNbdIOTimeout = "io-timeout"
)
var (
hasNBD = true
hasNBDCookieSupport = false
kernelCookieSupport = []util.KernelVersion{
{
Version: 5,
PatchLevel: 14,
SubLevel: 0,
ExtraVersion: 0,
Distribution: "",
Backport: false,
}, // standard 5.14+ versions
{
Version: 4,
PatchLevel: 18,
SubLevel: 0,
ExtraVersion: 365,
Distribution: ".el8",
Backport: true,
}, // CentOS-8.x
}
)
type deviceInfo interface {
GetName() string
GetPool() string
GetRadosNamespace() string
GetDevice() string
}
// rbdDeviceInfo strongly typed JSON spec for rbd device list output (of type krbd).
type rbdDeviceInfo struct {
// Name will only be set for krbd devices (NBD uses Image)
Name string `json:"name"`
// Image will only be set for NBD devices (krbd uses Name)
Image string `json:"image"`
Pool string `json:"pool"`
RadosNamespace string `json:"namespace"`
Device string `json:"device"`
}
func (rdi *rbdDeviceInfo) GetName() string {
if rdi.Name != "" {
return rdi.Name
}
return rdi.Image
}
func (rdi *rbdDeviceInfo) GetPool() string {
return rdi.Pool
}
func (rdi *rbdDeviceInfo) GetRadosNamespace() string {
return rdi.RadosNamespace
}
func (rdi *rbdDeviceInfo) GetDevice() string {
return rdi.Device
}
type detachRBDImageArgs struct {
imageOrDeviceSpec string
isImageSpec bool
isNbd bool
encrypted bool
volumeID string
unmapOptions string
logDir string
logStrategy string
}
// getDeviceList queries rbd about mapped devices and returns a list of deviceInfo
// It will selectively list devices mapped using krbd or nbd as specified by accessType.
func getDeviceList(ctx context.Context, accessType string) ([]deviceInfo, error) {
var (
stdout string
err error
)
if accessType == accessTypeKRbd {
// rbd device list --format json --device-type [krbd|nbd]
stdout, _, err = util.ExecCommand(ctx, rbd, "device", "list", "--format=json", "--device-type", accessType)
if err != nil {
return nil, fmt.Errorf("error getting device list from rbd for devices of type (%s): %w", accessType, err)
}
} else {
stdout, _, err = util.ExecCommand(ctx, "rbd-nbd", "list-mapped", "--format=json")
if err != nil {
return nil, fmt.Errorf("error getting device list from rbd-nbd for devices of type (%s): %w", accessType, err)
}
}
var devices []*rbdDeviceInfo
err = json.Unmarshal([]byte(stdout), &devices)
if err != nil {
return nil, fmt.Errorf(
"error to parse JSON output of device list for devices of type (%s): %w",
accessType,
err)
}
// convert []rbdDeviceInfo to []deviceInfo
deviceList := make([]deviceInfo, len(devices))
for i := range devices {
deviceList[i] = devices[i]
}
return deviceList, nil
}
// findDeviceMappingImage finds a devicePath, if available, based on image spec (pool/{namespace/}image) on the node.
func findDeviceMappingImage(ctx context.Context, pool, namespace, image string, useNbdDriver bool) (string, bool) {
accessType := accessTypeKRbd
if useNbdDriver {
accessType = accessTypeNbd
}
imageSpec := fmt.Sprintf("%s/%s", pool, image)
if namespace != "" {
imageSpec = fmt.Sprintf("%s/%s/%s", pool, namespace, image)
}
deviceList, err := getDeviceList(ctx, accessType)
if err != nil {
log.WarningLog(ctx, "failed to determine if image (%s) is mapped to a device (%v)", imageSpec, err)
return "", false
}
for _, device := range deviceList {
if device.GetName() == image && device.GetPool() == pool && device.GetRadosNamespace() == namespace {
return device.GetDevice(), true
}
}
return "", false
}
// Stat a path, if it doesn't exist, retry maxRetries times.
func waitForPath(ctx context.Context, pool, namespace, image string, maxRetries int, useNbdDriver bool) (string, bool) {
for i := range maxRetries {
if i != 0 {
time.Sleep(time.Second)
}
device, found := findDeviceMappingImage(ctx, pool, namespace, image, useNbdDriver)
if found {
return device, found
}
}
return "", false
}
// SetRbdNbdToolFeatures sets features available with rbd-nbd, and NBD module
// loaded status.
func SetRbdNbdToolFeatures() {
var stderr string
// check if the module is loaded or compiled in
_, err := os.Stat("/sys/module/" + moduleNbd)
if os.IsNotExist(err) {
// try to load the module
_, stderr, err = util.ExecCommand(context.TODO(), "modprobe", moduleNbd)
if err != nil {
hasNBD = false
log.WarningLogMsg("nbd modprobe failed (%v): %q", err, stderr)
return
}
}
log.DefaultLog("nbd module loaded")
// fetch the current running kernel info
release, err := util.GetKernelVersion()
if err != nil {
log.WarningLogMsg("fetching current kernel version failed (%v)", err)
return
}
if !util.CheckKernelSupport(release, kernelCookieSupport) {
log.WarningLogMsg("kernel version %q doesn't support cookie feature", release)
return
}
log.DefaultLog("kernel version %q supports cookie feature", release)
// check if the rbd-nbd tool supports cookie
stdout, stderr, err := util.ExecCommand(context.TODO(), rbdTonbd, "--help")
if err != nil || stderr != "" {
hasNBD = false
log.WarningLogMsg("running rbd-nbd --help failed with error:%v, stderr:%s", err, stderr)
return
}
if !strings.Contains(stdout, "--cookie") {
log.WarningLogMsg("rbd-nbd tool doesn't support cookie feature")
return
}
hasNBDCookieSupport = true
log.DefaultLog("rbd-nbd tool supports cookie feature")
}
// parseMapOptions helps parse formatted mapOptions and unmapOptions and
// returns mounter specific options.
func parseMapOptions(mapOptions string) (string, string, error) {
var krbdMapOptions, nbdMapOptions string
for _, item := range strings.Split(mapOptions, ";") {
var mounter, options string
if item == "" {
continue
}
s := strings.SplitN(item, ":", 2)
if len(s) == 1 {
options = strings.TrimSpace(s[0])
krbdMapOptions = options
} else {
// options might also contain values delimited with ":", in this
// case mounter type MUST be specified.
// ex: krbd:read_from_replica=localize,crush_location=zone:zone1;
mounter = strings.TrimSpace(s[0])
options = strings.TrimSpace(s[1])
switch strings.ToLower(mounter) {
case accessTypeKRbd:
krbdMapOptions = options
case accessTypeNbd:
nbdMapOptions = options
default:
return "", "", fmt.Errorf("unknown mounter type: %q, please specify mounter type", mounter)
}
}
}
return krbdMapOptions, nbdMapOptions, nil
}
// getMapOptions is a wrapper func, calls parse map/unmap funcs and feeds the
// rbdVolume object.
func (ns *NodeServer) getMapOptions(req *csi.NodeStageVolumeRequest, rv *rbdVolume) error {
krbdMapOptions, nbdMapOptions, err := parseMapOptions(req.GetVolumeContext()["mapOptions"])
if err != nil {
return err
}
krbdUnmapOptions, nbdUnmapOptions, err := parseMapOptions(req.GetVolumeContext()["unmapOptions"])
if err != nil {
return err
}
if rv.Mounter == rbdDefaultMounter {
rv.MapOptions = krbdMapOptions
rv.UnmapOptions = krbdUnmapOptions
} else if rv.Mounter == rbdNbdMounter {
rv.MapOptions = nbdMapOptions
rv.UnmapOptions = nbdUnmapOptions
}
readAffinityMapOptions, err := util.GetReadAffinityMapOptions(
util.CsiConfigFile, rv.ClusterID, ns.CLIReadAffinityOptions, ns.NodeLabels,
)
if err != nil {
return err
}
rv.appendReadAffinityMapOptions(readAffinityMapOptions)
return nil
}
func attachRBDImage(ctx context.Context, volOptions *rbdVolume, device string, cr *util.Credentials) (string, error) {
var err error
image := volOptions.RbdImageName
useNBD := false
if volOptions.Mounter == rbdTonbd && hasNBD {
useNBD = true
}
devicePath, found := waitForPath(ctx, volOptions.Pool, volOptions.RadosNamespace, image, 1, useNBD)
if !found {
backoff := wait.Backoff{
Duration: rbdImageWatcherInitDelay,
Factor: rbdImageWatcherFactor,
Steps: rbdImageWatcherSteps,
}
err = waitForrbdImage(ctx, backoff, volOptions)
if err != nil {
return "", err
}
devicePath, err = createPath(ctx, volOptions, device, cr)
}
return devicePath, err
}
func appendNbdDeviceTypeAndOptions(cmdArgs []string, userOptions, cookie string) []string {
isUnmap := slices.Contains(cmdArgs, "unmap")
if !isUnmap {
if !strings.Contains(userOptions, useNbdNetlink) {
cmdArgs = append(cmdArgs, "--"+useNbdNetlink)
}
if !strings.Contains(userOptions, setNbdReattach) {
cmdArgs = append(cmdArgs, fmt.Sprintf("--%s=%d", setNbdReattach, defaultNbdReAttachTimeout))
}
if !strings.Contains(userOptions, setNbdIOTimeout) {
cmdArgs = append(cmdArgs, fmt.Sprintf("--%s=%d", setNbdIOTimeout, defaultNbdIOTimeout))
}
rbd: utilize cookie support from rbd for nbd Problem: On remap/attach of device (i.e. nodeplugin restart), there is no way for rbd-nbd to defend if the backend storage is matching with the initial backend storage. Say, if an initial map request for backend "pool1/image1" got mapped to /dev/nbd0 and the userspace process is terminated (on nodeplugin restart). A next remap/attach (nodeplugin start) request within reattach-timeout is allowed to use /dev/nbd0 for a different backend "pool1/image2" For example, an operation like below could be dangerous: $ sudo rbd-nbd map --try-netlink rbd-pool/ext4-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="bfc444b4-64b1-418f-8b36-6e0d170cfc04" TYPE="ext4" $ sudo pkill -15 rbd-nbd <-- nodeplugin terminate $ sudo rbd-nbd attach --try-netlink --device /dev/nbd0 rbd-pool/xfs-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="d29bf343-6570-4069-a9ea-2fa156ced908" TYPE="xfs" Solution: rbd-nbd/kernel now provides a way to keep some metadata in sysfs to identify between the device and the backend, so that when a remap/attach request is made, rbd-nbd can compare and avoid such dangerous operations. With the provided solution, as part of the initial map request, backend cookie (ceph-csi VOLID) can be stored in the sysfs per device config, so that on a remap/attach request rbd-nbd will check and validate if the backend per device cookie matches with the initial map backend with the help of cookie. At Ceph-csi we use VOLID as device cookie, which will be unique, we pass the VOLID as cookie at map and use the same at the time of attach, that way rbd-nbd can identify backends and their matching devices. Requires: https://github.com/ceph/ceph/pull/41323 https://lkml.org/lkml/2021/4/29/274 Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
2021-09-21 13:13:10 +00:00
if hasNBDCookieSupport {
cmdArgs = append(cmdArgs, "--cookie="+cookie)
rbd: utilize cookie support from rbd for nbd Problem: On remap/attach of device (i.e. nodeplugin restart), there is no way for rbd-nbd to defend if the backend storage is matching with the initial backend storage. Say, if an initial map request for backend "pool1/image1" got mapped to /dev/nbd0 and the userspace process is terminated (on nodeplugin restart). A next remap/attach (nodeplugin start) request within reattach-timeout is allowed to use /dev/nbd0 for a different backend "pool1/image2" For example, an operation like below could be dangerous: $ sudo rbd-nbd map --try-netlink rbd-pool/ext4-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="bfc444b4-64b1-418f-8b36-6e0d170cfc04" TYPE="ext4" $ sudo pkill -15 rbd-nbd <-- nodeplugin terminate $ sudo rbd-nbd attach --try-netlink --device /dev/nbd0 rbd-pool/xfs-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="d29bf343-6570-4069-a9ea-2fa156ced908" TYPE="xfs" Solution: rbd-nbd/kernel now provides a way to keep some metadata in sysfs to identify between the device and the backend, so that when a remap/attach request is made, rbd-nbd can compare and avoid such dangerous operations. With the provided solution, as part of the initial map request, backend cookie (ceph-csi VOLID) can be stored in the sysfs per device config, so that on a remap/attach request rbd-nbd will check and validate if the backend per device cookie matches with the initial map backend with the help of cookie. At Ceph-csi we use VOLID as device cookie, which will be unique, we pass the VOLID as cookie at map and use the same at the time of attach, that way rbd-nbd can identify backends and their matching devices. Requires: https://github.com/ceph/ceph/pull/41323 https://lkml.org/lkml/2021/4/29/274 Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
2021-09-21 13:13:10 +00:00
}
}
if userOptions != "" {
// userOptions is appended after, possibly overriding the above
// default options.
cmdArgs = append(cmdArgs, "--"+userOptions)
}
return cmdArgs
}
func appendKRbdDeviceTypeAndOptions(cmdArgs []string, userOptions string) []string {
// Enable mapping and unmapping images from a non-initial network
// namespace (e.g. for Multus CNI). The network namespace must be
// owned by the initial user namespace.
cmdArgs = append(cmdArgs, "--device-type", accessTypeKRbd, "--options", "noudev")
if userOptions != "" {
// userOptions is appended after, possibly overriding the above
// default options.
cmdArgs = append(cmdArgs, "--options", userOptions)
}
return cmdArgs
}
// appendRbdNbdCliOptions append mandatory options and convert list of useroptions
// provided for rbd integrated cli to rbd-nbd cli format specific.
rbd: utilize cookie support from rbd for nbd Problem: On remap/attach of device (i.e. nodeplugin restart), there is no way for rbd-nbd to defend if the backend storage is matching with the initial backend storage. Say, if an initial map request for backend "pool1/image1" got mapped to /dev/nbd0 and the userspace process is terminated (on nodeplugin restart). A next remap/attach (nodeplugin start) request within reattach-timeout is allowed to use /dev/nbd0 for a different backend "pool1/image2" For example, an operation like below could be dangerous: $ sudo rbd-nbd map --try-netlink rbd-pool/ext4-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="bfc444b4-64b1-418f-8b36-6e0d170cfc04" TYPE="ext4" $ sudo pkill -15 rbd-nbd <-- nodeplugin terminate $ sudo rbd-nbd attach --try-netlink --device /dev/nbd0 rbd-pool/xfs-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="d29bf343-6570-4069-a9ea-2fa156ced908" TYPE="xfs" Solution: rbd-nbd/kernel now provides a way to keep some metadata in sysfs to identify between the device and the backend, so that when a remap/attach request is made, rbd-nbd can compare and avoid such dangerous operations. With the provided solution, as part of the initial map request, backend cookie (ceph-csi VOLID) can be stored in the sysfs per device config, so that on a remap/attach request rbd-nbd will check and validate if the backend per device cookie matches with the initial map backend with the help of cookie. At Ceph-csi we use VOLID as device cookie, which will be unique, we pass the VOLID as cookie at map and use the same at the time of attach, that way rbd-nbd can identify backends and their matching devices. Requires: https://github.com/ceph/ceph/pull/41323 https://lkml.org/lkml/2021/4/29/274 Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
2021-09-21 13:13:10 +00:00
func appendRbdNbdCliOptions(cmdArgs []string, userOptions, cookie string) []string {
if !strings.Contains(userOptions, useNbdNetlink) {
cmdArgs = append(cmdArgs, "--"+useNbdNetlink)
}
if !strings.Contains(userOptions, setNbdReattach) {
cmdArgs = append(cmdArgs, fmt.Sprintf("--%s=%d", setNbdReattach, defaultNbdReAttachTimeout))
}
if !strings.Contains(userOptions, setNbdIOTimeout) {
cmdArgs = append(cmdArgs, fmt.Sprintf("--%s=%d", setNbdIOTimeout, defaultNbdIOTimeout))
}
rbd: utilize cookie support from rbd for nbd Problem: On remap/attach of device (i.e. nodeplugin restart), there is no way for rbd-nbd to defend if the backend storage is matching with the initial backend storage. Say, if an initial map request for backend "pool1/image1" got mapped to /dev/nbd0 and the userspace process is terminated (on nodeplugin restart). A next remap/attach (nodeplugin start) request within reattach-timeout is allowed to use /dev/nbd0 for a different backend "pool1/image2" For example, an operation like below could be dangerous: $ sudo rbd-nbd map --try-netlink rbd-pool/ext4-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="bfc444b4-64b1-418f-8b36-6e0d170cfc04" TYPE="ext4" $ sudo pkill -15 rbd-nbd <-- nodeplugin terminate $ sudo rbd-nbd attach --try-netlink --device /dev/nbd0 rbd-pool/xfs-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="d29bf343-6570-4069-a9ea-2fa156ced908" TYPE="xfs" Solution: rbd-nbd/kernel now provides a way to keep some metadata in sysfs to identify between the device and the backend, so that when a remap/attach request is made, rbd-nbd can compare and avoid such dangerous operations. With the provided solution, as part of the initial map request, backend cookie (ceph-csi VOLID) can be stored in the sysfs per device config, so that on a remap/attach request rbd-nbd will check and validate if the backend per device cookie matches with the initial map backend with the help of cookie. At Ceph-csi we use VOLID as device cookie, which will be unique, we pass the VOLID as cookie at map and use the same at the time of attach, that way rbd-nbd can identify backends and their matching devices. Requires: https://github.com/ceph/ceph/pull/41323 https://lkml.org/lkml/2021/4/29/274 Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
2021-09-21 13:13:10 +00:00
if hasNBDCookieSupport {
cmdArgs = append(cmdArgs, "--cookie="+cookie)
rbd: utilize cookie support from rbd for nbd Problem: On remap/attach of device (i.e. nodeplugin restart), there is no way for rbd-nbd to defend if the backend storage is matching with the initial backend storage. Say, if an initial map request for backend "pool1/image1" got mapped to /dev/nbd0 and the userspace process is terminated (on nodeplugin restart). A next remap/attach (nodeplugin start) request within reattach-timeout is allowed to use /dev/nbd0 for a different backend "pool1/image2" For example, an operation like below could be dangerous: $ sudo rbd-nbd map --try-netlink rbd-pool/ext4-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="bfc444b4-64b1-418f-8b36-6e0d170cfc04" TYPE="ext4" $ sudo pkill -15 rbd-nbd <-- nodeplugin terminate $ sudo rbd-nbd attach --try-netlink --device /dev/nbd0 rbd-pool/xfs-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="d29bf343-6570-4069-a9ea-2fa156ced908" TYPE="xfs" Solution: rbd-nbd/kernel now provides a way to keep some metadata in sysfs to identify between the device and the backend, so that when a remap/attach request is made, rbd-nbd can compare and avoid such dangerous operations. With the provided solution, as part of the initial map request, backend cookie (ceph-csi VOLID) can be stored in the sysfs per device config, so that on a remap/attach request rbd-nbd will check and validate if the backend per device cookie matches with the initial map backend with the help of cookie. At Ceph-csi we use VOLID as device cookie, which will be unique, we pass the VOLID as cookie at map and use the same at the time of attach, that way rbd-nbd can identify backends and their matching devices. Requires: https://github.com/ceph/ceph/pull/41323 https://lkml.org/lkml/2021/4/29/274 Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
2021-09-21 13:13:10 +00:00
}
if userOptions != "" {
options := strings.Split(userOptions, ",")
for _, opt := range options {
cmdArgs = append(cmdArgs, "--"+opt)
}
}
return cmdArgs
}
func createPath(ctx context.Context, volOpt *rbdVolume, device string, cr *util.Credentials) (string, error) {
isNbd := false
imagePath := volOpt.String()
log.TraceLog(ctx, "rbd: map mon %s", volOpt.Monitors)
mapArgs := []string{
"--id", cr.ID,
"-m", volOpt.Monitors,
"--keyfile=" + cr.KeyFile,
}
// Choose access protocol
if volOpt.Mounter == rbdTonbd && hasNBD {
isNbd = true
}
cli := rbd
if isNbd {
cli = rbdNbdMounter
mapArgs = append(mapArgs, "--log-file",
getCephClientLogFileName(volOpt.VolID, volOpt.LogDir, "rbd-nbd"))
}
if device != "" {
// TODO: use rbd cli for attach/detach in the future
cli = rbdNbdMounter
mapArgs = append(mapArgs, "attach", imagePath, "--device", device)
rbd: utilize cookie support from rbd for nbd Problem: On remap/attach of device (i.e. nodeplugin restart), there is no way for rbd-nbd to defend if the backend storage is matching with the initial backend storage. Say, if an initial map request for backend "pool1/image1" got mapped to /dev/nbd0 and the userspace process is terminated (on nodeplugin restart). A next remap/attach (nodeplugin start) request within reattach-timeout is allowed to use /dev/nbd0 for a different backend "pool1/image2" For example, an operation like below could be dangerous: $ sudo rbd-nbd map --try-netlink rbd-pool/ext4-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="bfc444b4-64b1-418f-8b36-6e0d170cfc04" TYPE="ext4" $ sudo pkill -15 rbd-nbd <-- nodeplugin terminate $ sudo rbd-nbd attach --try-netlink --device /dev/nbd0 rbd-pool/xfs-image /dev/nbd0 $ sudo blkid /dev/nbd0 /dev/nbd0: UUID="d29bf343-6570-4069-a9ea-2fa156ced908" TYPE="xfs" Solution: rbd-nbd/kernel now provides a way to keep some metadata in sysfs to identify between the device and the backend, so that when a remap/attach request is made, rbd-nbd can compare and avoid such dangerous operations. With the provided solution, as part of the initial map request, backend cookie (ceph-csi VOLID) can be stored in the sysfs per device config, so that on a remap/attach request rbd-nbd will check and validate if the backend per device cookie matches with the initial map backend with the help of cookie. At Ceph-csi we use VOLID as device cookie, which will be unique, we pass the VOLID as cookie at map and use the same at the time of attach, that way rbd-nbd can identify backends and their matching devices. Requires: https://github.com/ceph/ceph/pull/41323 https://lkml.org/lkml/2021/4/29/274 Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
2021-09-21 13:13:10 +00:00
mapArgs = appendRbdNbdCliOptions(mapArgs, volOpt.MapOptions, volOpt.VolID)
} else {
mapArgs = append(mapArgs, "map", imagePath)
if isNbd {
mapArgs = appendNbdDeviceTypeAndOptions(mapArgs, volOpt.MapOptions, volOpt.VolID)
} else {
mapArgs = appendKRbdDeviceTypeAndOptions(mapArgs, volOpt.MapOptions)
}
}
if volOpt.readOnly {
mapArgs = append(mapArgs, "--read-only")
}
var (
stdout string
stderr string
err error
)
if volOpt.NetNamespaceFilePath != "" {
stdout, stderr, err = util.ExecuteCommandWithNSEnter(ctx, volOpt.NetNamespaceFilePath, cli, mapArgs...)
} else {
stdout, stderr, err = util.ExecCommand(ctx, cli, mapArgs...)
}
if err != nil {
log.WarningLog(ctx, "rbd: map error %v, rbd output: %s", err, stderr)
// unmap rbd image if connection timeout
if strings.Contains(err.Error(), rbdMapConnectionTimeout) {
dArgs := detachRBDImageArgs{
imageOrDeviceSpec: imagePath,
isImageSpec: true,
isNbd: isNbd,
encrypted: volOpt.isBlockEncrypted(),
volumeID: volOpt.VolID,
unmapOptions: volOpt.UnmapOptions,
logDir: volOpt.LogDir,
logStrategy: volOpt.LogStrategy,
}
detErr := detachRBDImageOrDeviceSpec(ctx, &dArgs)
if detErr != nil {
log.WarningLog(ctx, "rbd: %s unmap error %v", imagePath, detErr)
}
}
return "", fmt.Errorf("rbd: map failed with error %w, rbd error output: %s", err, stderr)
}
devicePath := strings.TrimSuffix(stdout, "\n")
return devicePath, nil
}
func waitForrbdImage(ctx context.Context, backoff wait.Backoff, volOptions *rbdVolume) error {
imagePath := volOptions.String()
err := wait.ExponentialBackoff(backoff, func() (bool, error) {
used, err := volOptions.isInUse()
if err != nil {
return false, fmt.Errorf("fail to check rbd image status: (%w)", err)
}
if (volOptions.DisableInUseChecks) && (used) {
log.UsefulLog(ctx, "valid multi-node attach requested, ignoring watcher in-use result")
return used, nil
}
return !used, nil
})
// return error if rbd image has not become available for the specified timeout
if wait.Interrupted(err) {
return fmt.Errorf("rbd image %s is still being used", imagePath)
}
// return error if any other errors were encountered during waiting for the image to become available
return err
}
func detachRBDDevice(ctx context.Context, devicePath, volumeID, unmapOptions string, encrypted bool) error {
nbdType := false
if strings.HasPrefix(devicePath, "/dev/nbd") {
nbdType = true
}
dArgs := detachRBDImageArgs{
imageOrDeviceSpec: devicePath,
isImageSpec: false,
isNbd: nbdType,
encrypted: encrypted,
volumeID: volumeID,
unmapOptions: unmapOptions,
}
return detachRBDImageOrDeviceSpec(ctx, &dArgs)
}
// detachRBDImageOrDeviceSpec detaches an rbd imageSpec or devicePath, with additional checking
// when imageSpec is used to decide if image is already unmapped.
func detachRBDImageOrDeviceSpec(
ctx context.Context,
dArgs *detachRBDImageArgs,
) error {
if dArgs.encrypted {
mapperFile, mapperPath := util.VolumeMapper(dArgs.volumeID)
mappedDevice, mapper, err := util.DeviceEncryptionStatus(ctx, mapperPath)
if err != nil {
log.ErrorLog(ctx, "error determining LUKS device on %s, %s: %s",
mapperPath, dArgs.imageOrDeviceSpec, err)
return err
}
if mapper != "" {
// mapper found, so it is open Luks device
err = util.CloseEncryptedVolume(ctx, mapperFile)
if err != nil {
log.ErrorLog(ctx, "error closing LUKS device on %s, %s: %s",
mapperPath, dArgs.imageOrDeviceSpec, err)
return err
}
dArgs.imageOrDeviceSpec = mappedDevice
}
}
unmapArgs := []string{"unmap", dArgs.imageOrDeviceSpec}
if dArgs.isNbd {
unmapArgs = appendNbdDeviceTypeAndOptions(unmapArgs, dArgs.unmapOptions, dArgs.volumeID)
} else {
unmapArgs = appendKRbdDeviceTypeAndOptions(unmapArgs, dArgs.unmapOptions)
}
var err error
var stderr string
if dArgs.isNbd {
_, stderr, err = util.ExecCommand(ctx, rbdTonbd, unmapArgs...)
} else {
_, stderr, err = util.ExecCommand(ctx, rbd, unmapArgs...)
}
if err != nil {
// Messages for krbd and nbd differ, hence checking either of them for missing mapping
// This is not applicable when a device path is passed in
if dArgs.isImageSpec &&
(strings.Contains(stderr, fmt.Sprintf(rbdUnmapCmdkRbdMissingMap, dArgs.imageOrDeviceSpec)) ||
strings.Contains(stderr, fmt.Sprintf(rbdUnmapCmdNbdMissingMap, dArgs.imageOrDeviceSpec))) {
// Devices found not to be mapped are treated as a successful detach
log.TraceLog(ctx, "image or device spec (%s) not mapped", dArgs.imageOrDeviceSpec)
return nil
}
return fmt.Errorf("rbd: unmap for spec (%s) failed (%w): (%s)", dArgs.imageOrDeviceSpec, err, stderr)
}
if dArgs.isNbd && dArgs.logDir != "" {
logFile := getCephClientLogFileName(dArgs.volumeID, dArgs.logDir, "rbd-nbd")
go strategicActionOnLogFile(ctx, dArgs.logStrategy, logFile)
}
return nil
}